Archive for September, 2010

Classifying aquifers

Tuesday, September 28th, 2010

October 2010 Article: An Aquifer Classification System and Geographical Information System-Based Analysis Tool for Watershed Managers in the Western U.S., by Scott M. Payne and William W. Woessner.

Example of aquifer mapping

I wish I had a dollar for every classification system ever proposed. This one, however, got my attention because it seems widely applicable, repeatable, and reduces sometimes cumbersome complex databases and analyzes to straightforward terminology and graphical representations. Moreover, it’s based on a watershed scale.

The proposed classification system uses basin geology, aquifer productivity, water quality, and the degree of groundwater/surface water connection as classification criteria. The approach is based on literature values, reference databases, and fundamental hydrologic and hydrogeologic principles. The proposed classification system treats dataset completeness as a variable and includes a tiered assessment protocol that depends on the quality and quantity of data. The hierarchical approach is designed to improve communication between groundwater professionals and natural resource managers, similar to the classification system for natural rivers developed by Rosgen.

Classification systems always seem to involve different opinions over how things should be lumped, split, summarized, inspected, detected, neglected, and selected, and this one is no exception. The paper went through three full rounds of reviews before reaching tentative acceptance. Many thanks go to the reviewers who offered so many helpful comments.

[Please note: I have quoted and paraphrased freely from the article, but the interpretation is my own!]

News coverage

Monday, September 20th, 2010

The Hirsch et al article in the October 2010 issue (see previous postings) has garnered some news coverage. Looking at the different articles, it’s interesting to see how the news media perceived the news release from USGS and the JAWRA article. Some note the new methodology, like I did in my posting here, while others focus on the findings relevant to Chesapeake Bay.

Here are some of the news articles I found in a Google search today:

News Release on Chesapeake Bay

Wednesday, September 15th, 2010

USGS has issued a news release concerning the Hirsch et al. paper now available in EarlyView. The release focuses on the findings presented in the paper concerning Chesapeake Bay.

Time, Discharge, and Season

Sunday, September 12th, 2010

October 2010 Article: Weighted Regressions on Time, Discharge, and Season (WRTDS), with an Application to Chesapeake Bay River Inputs, by Robert M. Hirsch, Douglas L. Moyer, Stacey A. Archfield. (Available OnlineOpen.)

This article marks the return of Bob Hirsch to publishing real science after a 20-year hiatus in senior management. Bob and I talked about this paper last January. Recognizing its importance, I agreed to expedite publication production, with the condition the review process had to proceed normally. The paper likely will have a significant effect on how we look at water-quality trends in the future.

The paper presents a new approach to the analysis of long-term surface water-quality data weighted regressions of concentrations on time, discharge, and season. The method is formulated to allow for maximum flexibility in representations of the long-term trend, seasonal components, and discharge-related components of the behavior of the water-quality variable of interest.

The method’s conceptual roots are recognizable to anyone familiar with Locally Weighted Scatterplot Smoothing (LOWESS) for flow-adjusting concentrations. Think of this as LOWESS in multiple dimensions. It’s very computer intensive and requires long time series of data, but the authors point out these no longer pose the limitations they did 20 years ago.

The authors apply their method to datasets for the nine large tributaries of Chesapeake Bay from 1978 to 2008. Although only a preliminary analysis, the outputs are intended to be useful not only as measures of relative success for achievement of water-quality goals but also as a gateway to diagnostic analysis of the nature of and possible reasons for the changes that are being observed.

[Please note: I have quoted and paraphrased freely from the article, but the interpretation is my own!]

Data Citations

Thursday, September 2nd, 2010

The 24 August 2010 issue of EOS ( — You need password access.) contains an interesting article,  ”Data Citations and Peer Review” by Mark. A. Parsons, Ruth Duerr, and J.-B. Minster. They note, “Ultimately, more is needed to develop completely unambiguous ways to cite data precisely … journal editors and reviewers would need to be more rigorous in demanding that authors accurately cite the data they use in their research.” Amen to that — see my posting of 11 September 2009.

JAWRA editorial policy differs from their recommendations in one respect. Last year, in revising our Instructions for Authors, we basically threw up our hands on how to cite data sets. Who is the “author” of data compiled in three centuries? What is the meaning of year cited (e.g., USGS 1997) in a the traditional citation format, when minutes may count? We felt treating data sets as informal references, not Literature Cited, would give authors a lot more freedom to focus on the main point of clearly identifying data that were actually used.

Downloads, rather than formal citations, are perhaps a better metric for big data providers like USGS’ NWIS, and EPA’s STORET, which are accessed by thousands of private citizens. Also, GIS data today are managed in data bases like the National Hydrography Dataset. The notion of citing a small, self contained GIS “coverage” prepared by a single author almost seems quaint and old fashioned. (I’m dating myself even admitting I know about coverages!)

Nevertheless, I’m willing to admit our approach may short-change some data compilers looking for the credit of formal citations. And, I completely agree careful compilation and editing of data sets is something which needs to be professionally recognized.

The answer may be allowing formal citations of data sets where appropriate. Exactly what is appropriate is where I need help. Your comments on this matter are invited.